J. Am. Chem. SOC.1987, 109, 613-615
Figure 1. X-ray structure of 2a. Hydrogen atoms have been omitted for clarity. Selected bond distances (in angstroms): C1-02 = 1.378 (2), C3-02 = 1.478 (2),CI-07 = 1.434(2). 06-07 = 1.465 (2), 06-C5 = 1.449 (2). Bond angles: C1-02-C3 = 119.9 (l)', C18-C3-C22 = 117.9 (1)'.
613
reaction mixture led to ketone 3b as the only isolable product. Apparently, replacement of a tert-butyl group by methyl allows greater conformational flexibility in the carbonyl oxide 4b, so that intramolecular cyclization is no longer favored. Indeed, rotation about the ester alkyl oxygen bond (toward R = CH, in Figure 2) moves the ester group out of alignment with the carbonyl oxide, so that intermolecular reaction of the carbonyl oxide with, for example, the cognate formaldehyde, can compete successfully. Preliminary studies indicate a strong electronic influence on the cyclization. Ozonolysis of benzoate ester ICgave quite cleanly the ketone 3c-no ozonide product was detected. Since 4c is subject to the same steric constraints as 4a, the change in products reflects the decreased dipolarophilicity of the benzoate carbonyl group as compared to the p-nitrobenzoate.' Activation by an electron-withdrawing group is again manifested in Id, which leads to a 1:l mixture of 2d4 and 3d. The reaction of trifluoroacetate l e gives a complex mixture of products, from which 2e4 can be isolated in 18% yield. Thus, these early experiments establish rather strict steric and electronic requirements for the intramolecular carbonyl oxide-ester cycloaddition. Acknowledgment. W.B. thanks the donors of the Petroleum Research Fund, administered by the American Chemical Society, for the support of this work. Supplementary Material Available: Tables of positional parameters, thermal parameters, interatomic distances, interatomic angles, and dihedral angles for non-hydrogen atoms (7 pages). Ordering information is given on any current masthead page.
\
t0u
4
(7)Electron-withdrawing groups are known to increase the dipolarophilicity of carbonyl groups toward carbonyl oxides. See ref 2.
Figure 2. Conformation of 4, viewed in Newman projection along the ester oxygen-alkyl bond. When R = Me, a clockwise rotation about this bond is favorable, removing the ester C=O from proximity to the carbonyl oxide.
the dioxane ring of 2a exists in an envelope conformation with 0 8 lying 0.81 A out of the least-squares plane calculated for C1-02-C3-C4-C5. The flattened ring minimizes transannular interactions of the gem-di-tert-butyl system with the epoxy or peroxy bridges. The five-membered ring adopts a half-chair conformation with C5 situated 0.26 8, below, and 0 8 0.46 8, above, the C1-07-06 plane. At C1, the axial peroxy bridge sets up an anomeric effect with 0 2 . Indeed, strong differentiation of bond lengths to this center (Cl-02, 1.378 (2) A, C3-02, 1.478 (2) A) confirm this interaction, as established for alkyl pyranosides.6 The formation of 2a is considered to involve cycloaddition of the carbonyl oxide moiety with the ester group within 4a (Figure 2), which is produced on fragmentation of the primary ozonide of la. Thus, ozonolysis of la, followed by addition of dimethyl sulfide at -70 OC, leads to a 1:l mixture of 2a and 3a in essentially quantitative yield. Since 2a does not react with dimethyl sulfide, this result requires reductive diversion of a percursor of 2a to give 3a. From the Criegee mechanism, only the carbonyl oxide 4a satisfies this role. Examination of molecular models indicates that only the syn geometry of carbonyl oxide 4 is suitably disposed for intramolecular cycloaddition. Moreover, it is clear that the steric effect of the gem-di-tert-butyl segment operates to enforce a conformation of 4a which places the reactive groups in close proximity. In this light, it was of interest to examine a less sterically biased analogue. Ozonolysis of l b gave none of the ozonide 2b. Instead, subsequent reduction (Me#) of the crude (5) Data were collected from a 0.3 X 0.3 X 0.4mm crystal with an Enraf-Nonius CAD4 diffractometer. Ci9H2,N06crystallizes from methanol in space group P2,/c,a = 7.624 ( I ) A, b = 12.778(3) A, c = 19.210(7) A, p = 92.50 (2)'. V = 1869 (2)A3, z = 4,daM = 1.296.The structure was solved (1881 reflections, I > 20) by direct methods and refined to R = 0.042 (R, = 0.056).
(6)(a) Jeffrey, G.A.; Pople, J. A.; Binkley, J. S.;Vishveshwara, S.J . Am. Chem. SOC.1978,100,373. (b) Moews, P. C.; Knox, J. R. J. Am. Chem. SOC. 1976,98, 6628.
An Example of Axial Selectivity in Nucleophilic Additions to Cyclohexanones and Cyclohexenones Barry M. Trost,* Josefa Florez, and David J. Jebaratnam Department of Chemistry, University of Wisconsin Madison, Wisconsin 53706 Received September 1 I , 1986 Controlling the stereochemistry of additions of carbanions to cyclic ketones remains an important contemporary problem. The concept that steric effects control such reactions leads to the normal introduction of nucleophiles in an equatorial orientation in the case of six-membered ring ketones.'** A recent interesting variation using a bulky aluminum alkyl directed the nucleophile in an axial fashion due to the preference to place the even bulkier aluminum alkoxide e q ~ a t o r i a l l y . ~On the other hand, orbital distortion arguments4 and consideration of torsional effectsS associated with the addition suggest that there may exist an intrinsic bias for axial attack-a bias which is frequently overwhelmed by ( I ) For some recent examples of enhancing equatorial attack, see: Mac Donald, T. L.; Still, W. C. J. Am. Chem. SOC.1975,97, 5280. Ashby, E.C.; Noding, S.A. J . Org. Chem. 1979,44, 4371. Weidmann, B.; Maycock, C. D.; Seebach, D. Helu. Chim. Acra 1981,64,1552. Reetz, M. T.Top. Curr. Chem. 1982,106,I . (2)For an exception in the case of a 1,3-dioxan-5-one, see: Kobayashi, Y. M.; Lambrecht, J.; Jochims, J. C.; Burkert, U. Chem. Ber. 1978,Ill, 3442. (3) Maruoka, K.; Itoh, T.; Yamamoto, H . J . Am. Chem. SOC.1985,107,
4573. (4) Klein, J. Tetrahedron Lett. 1973,4307; Tetrahedron, 1974,30,3349; Burgess, E. M.; Liotta, C. L. J . Org. Chem. 1981,46, 1703. Cieplak, A. S. J . Am. Chem. SOC. 1981,103,4540.Anh, N. T. Top. Curr. Chem. 1980,88, 145. (5) Cherest, M.; Felkin, H . Tetrahedron Lett. 1968,2205. Cherest, M. Tetrahedron 1980,36, 1593. Wu, Y. D.; Houk, K. N., unpublished results. We thank Professor Houk for sending a preprint prior to publication.
0002-7863 I87 I1 509-06 ~ 13x01 .SO10 ._ . ., .. . 0 1987 American Chemical Society I
,
614 J. A m . Chem. SOC.,Vol. 109, No. 2, 1987
Communications to the Editor
Table I. Effect of Metal Counterion on Diastereofacial Selectivity" Axial/Eauatoria!. Attack
2
yield
3
yield
5.5:1 5.5:l 1.8: 1 2.5:l 1:2.3
87 72 81 66 89
23:l 20:l 6.8:l
94 95 79
1.6: 1
67
metal Li+ K+C
MgCd Zn+ e AP+/
"All reactions performed in T H F at -78 OC unless otherwise noted. Generated by treating acetonitrile with lithium hexamethyldisilazide. Generated by treating acetonitrile with potassium hexamethyldisilazide. "Generated by addition of magnesium bromide to the lithium derivative. 'Generated by addition of zinc chloride to the lithium derivative. Reaction allowed to warm to room temperature before quenching. /Generated by adding methylaluminum bis(2,6-di-tert-butyl-4-phenoxide) to the lithium derivative.
nonbonded interactions. A corollary to this argument is that if a sufficiently nonsterically demanding nucleophile can be found, axial attack may become the preferred course.6 In conjunction with a natural product synthesis, we required addition of an acetaldehyde fragment onto a cyclohexenone in an axial fashion. We wish to report that additions of acetonitrile anion' to both cyclohexanones and cyclohexenones is generally highly axial selective. Carvone represented the pertinent case for our ultimate synthetic objective and was therefore examined first. A single adduct was obtained which subsequent correlation allowed us to assign as the product from axial attack, Le., 1. In order to determine
- Ho.Y-
o& LiCH2CN b< -a< +
NC
1
dccrease axial attack. In contrast to the work of Yamamoto with nonstabilized anions, employment of methylaluminum bis(2,6di-tert-butyl-4-methylphenoxide)increases the propensity for equatorial attack. The effective steric bulk of the reagent may account for these observations. Theoretical studies suggest that the lithium and sodium (and, by inference, potassium) salts of acetonitrile both prefer to place the metal close to nitrogen in a partially bridged bent ketenimine structure-then minimizing the effective steric bulk of the attacking nucleophile.I0 The more covalent metals such as magnesium appear to increase the covalency with carbon. The resulting increased steric demands of the attacking nucleophile then lead to enhanced nonbonded interactions between the reaction partners with the result of axial attack becoming less favored. Of course, aggregation effects as a result of the metal must not be ignored." I3C N M R data also proves useful in assigning stereochemistry. The signal for the methylene group of the acetonitrile side chain appears at higher field for the axial isomer than for the equatorial isomer due to steric compression; in addition, the carbinol carbon appears at lower fields for an equatorial hydroxyl group (see Table II).'* Similar results are obtained for the saturated-unsaturated ketone pair 8 and 9. Using the potassium anion derived from acetonitrile, the saturated ketone 8 gives a 6:l ratio of 109:11,9 and the unsaturated ketone 9 gives a 1O:l ratio of 129:13.9 NC
&77w L
whether the effect was associated with an enone as has been suggested in other cases,6d,gthe saturated and unsaturated pair of six-membered ring ketones 2 and 3 was examined. The saturated ketone forms the two adducts 4 and 5 in a 5.5:l ratio (as
5
determined by capillary VPC) with either the potassium or the lithium derivatives (see Table I) in excellent yields.*s9 Assignment of the stereochemistry depicted in 4 for the major adduct derives from the single-crystal X-ray analysis. The enone 3 also reacts smoothly to give 69 and 79 in a ratio of 20-23:l (Table I). Catalytic hydrogenation (3 atm of H2, PtO,, HOAc, room temperature) converts adduct 6 into 4-establishing that both have the same stereochemistry. Thus, the preference for axial attack is independent of the unsaturated nature ofthe ketone although the magnitude of the preference is enhanced in going from the saturated to the unsaturated acceptor. An examination of Table I reveals that the effect of metal on the stereochemistry may be profound. More covalent metals
'
OH
-p=\+, OH
Hoy=G&o,
P I
I"
0
NC'
?
U--
'?
'3
'b
Examination of the menthone (14)-pulegone (15) pair reveals the limitations to the preference for axial attack. Menthone reacts with lithiated acetonitrile to give a 20:l ratio of the two adduct^;^ pulegone under the same conditions gives a 22:l ratio of its add u c t ~ .Examination ~ of the I3C data for the menthone adducts indicates the major isomer is the product of equatorial attack, 17, not of axial attack, 16, whereas, the same data support the preference for axial attack in the case of pulegone. Verification
I
OH
that the major isomer from pulegone corresponds to the minor isomer from menthone arises by catalytic hydrogenation (H2,PtO,, 3 atm, HOAc, room temperature). Confirmation of the assignments of 16 and 17 is also obtained by conversion of the nitriles to the known esters.14 In this way, the major product from
Table 11. Relevant I3C Shifts for Stereochemical Assignment" axial ketone
CH,CN
2 8
21.3 28.2 24.2
14
HO-C-CH,CN
"See ref 13. b N o t detected.
unsaturated ketone adducts
saturated ketone adducts equatorial 70.8 71.2 73.4
CH,CN 33.0 32.8 30.2
HO-C-CH,CN 31.6 69.9 73.8
equatorial
axial ketone
3 9 15
CH,CN 30.2 30.2 28.7
HO-C-CH2CN 69.0 69.1 75.2
CH,CN 31.6 31.4
N D6
HO-C-CH2CN 61.3 67.6 12.3
J. A m . Chem. Sot. 1987, 109, 6 15-6 17
menthone is assigned as the equatorial isomer 17 and that from pulegone as the axial isomer 18. The change in stereochemistry of attack on menthone presumably derives from the combinatorial effects of the 2-isopropyl substituent and the @-axialhydrogens making the nonbonded interactions dominate over the intrinsic bias for axial attack. We attribute the ability of the acetonitrile salts to reflect the intrinsic bias of six-membered ring ketones to suffer axial attack to the small steric bulk associated with the attacking end of these ketenimine-type structures. Indeed, the decreasing order of axial selectivity in going from Li' to Mg2' agrees with the recent calculations suggesting greater bonding of Mg2+ to the carbon, thereby increasing its effective steric bulk. Furthermore, the results also seem to be more in accord with torsional effects rather then orbital distortion as the major contributor to the intrinsic axial bias in nucleophilic addition to six-membered ring ketones. In contrast to the results of Yamamoto, use of the bulky aluminum reagent decreased the axial selectivity. The difference may reside in the fact that our nucleophile is a stabilized anion, whereas his were not. The versatility of the nitrile for further structural elaboration makes this method for formation of an axial C-C bond quite synthetically useful. For example, DIBAL-H reduction converts the acetonitrile side chain to an axial acetaldehyde side chain which cannot derive from a crossed aldol condensation. While, at first glance, the case of menthone seems to be a limitation, synthetically, the axial isomer is available by the expedient of employing the unsaturated analogue followed by hydrogenation of its adduct 18 to give the axial isomer 16. The effect of an a-substituent to disfavor axial attack needs not dominate. In the case of the substrate 20 which bears an a-substituent that hinders axial attack, axial attack still dominated. It appears that by designing an appropriate nucleophile, axial rather than equatorial attack may be more generally available.
2P
2?
Acknowledgment. We thank the General Medical Sciences Institute of the National Institutes of Health for their generous support of our program. Josepha Florez is a Spanish Ministry of Education and Science Postdoctoral Fellow. (6) Acetylide additions appear to be axial selective. (a) Hennion, G. F.; O'Shea, F. X. J . Am. Chem. SOC.1958, 80, 614. (b) Akhrem, A. A,; Kamernitskii, A. V.; Prohoda, A. M. Zh. Org. Kfiim. 1967, 3, 50. (c) Rocquet, F.; Battioni, J. P.; Capmau, M.-L.; Chodkiewicz, W. C.R. Hebd. Seances Acad. Sci., Ser. C 1969,268, 1449. (d) Stork, G.; Stryker, J. M. Tetrahedron Lett. 1983, 24, 4887. (e) Kuwajima, I.; Nakamura, E.; Hashimoto, K. Terrahedron 1983, 39, 975. (f) Fleming, I.; Terrett, N. K. J . Organomet. Chem. 1984, 264, 99. (7) A 2:l ratio for addition of lithiated acetonitrile to 4-rerr-butylcyclohexanone in pentane-ether or PhH-HMPA has been reported. Bellassoued, M.; Dardoize, F.; Gaudemar-Bardone, F.; Gaudemar, M.; Goasdoue, N . Tetrahedron 1979, 32, 2713. Idriss, N.; Perry, M.; Maroni-Barnaud, Y.; Roux-Schmitt, M.-C.; Seyden-Penne, J. J . Cfiem. Res. Synop. 1978, 128. (8) The much higher axial selectivity we observe than previously found may be due to the difference in solvent. Also we find that there is a deterioration of axial selectivity with increased reaction times (normally reactions are complete within 15-30 min) which apparently arises by the addition being somewhat reversible. Such a factor may also be responsible for the lower selectivity of the above report. (9) All new compounds have been fully characterized and elemental composition established by combustion analysis and/or mass spectroscopy. (IO) Kaneti, J.; Schleyer, P. v. R.; Clark, T.; Kos, A. J.; Spitznagel, G. W.; Andrade, J. G.; Moffat, J. B. J . Am. Chem. SOC.1986, 108, 1481. Also see: Kruger, C. J . Organomet. Cfiem. 1967, 9, 125. Boche, G.; Marech, M.; Harms, K. Angew. Chem., Int. Ed. Engl. 1986, 25, 373. (11) Das, R.; Wilkie, C. A. J . Am. Chem. SOC.1972, 94, 4555. Bauer, W.; Seebach, D. Helu. Chim. Acta 1984, 67, 1972. A correction of the state of aggregation of LiCH2CN in T H F to 2 has been reported as a private communication (see ref 11 in the Kaneti, J., et al. paper, ref IO). (1 2) Stothers, J. B. Carbon-13 NMR Spectroscopy; Academic Press: New York, 1972. Senda, Y.; Ishiyama, J.; Imaizumi, S . Tetrahedron 1975, 31, 1601. (1 3) Assignment of the I3C signals was facilitated by preparing the adducts with KCD,CN. (14) Pansard, J.; Gaudemar, M. Bull. SOC.Chim. Fr. 1973, 12, 3472.
0002-7863/87/1509-0615$01.50/0
615
+
[6 31 Cycloaddition to Nine-Membered Ring Carbocycles Barry M. Trost* and Peter R. Seoane Department of Chemistry, University o j Wisconsin Madison. Wisconsin 53706 Received July 29, 1986
The Diels-Alder reaction continues to prove itself as the most important six-membered ring-forming reaction. In searching for alternative cycloaddition strategies for ring construction, we hoped that any such methods developed would be applicable to a variety of ring sizes. FMO theory suggests that tropones should preferentially react at C(2) and C(7) rather than C(2) and C(3).',* Indeed, in cycloadditions of dienes with these acceptors, a [6 + 41 pathway may compete with and/or dominate the [4 + 21 r e a ~ t i o n . We ~ wish to record our preliminary observations that demonstrate that 2- [ (trimethylsilyl)methyl]allyl carboxylates4 and their substituted analogues5 not only can form five-membered rings6 but undergo exclusive [6 31 nine-membered carbocycle formation in their reactions with tropones according to eq 1. To our knowledge, this reaction is the first report of a [6 31 cycloaddition.
+
+
0
I1
7'
OCR"
SiMe3
R
1
2
I
R
3
In order to test the feasibility of [6 + 31 cycloaddition of 2, R' = H and R" = CH,, with tropone, we reacted a 1:l mixture of the two in toluene at 80 OC using 2.5 mol % of a Pd(0) catalyst generated in situ by mixing palladium acetate and triisopropyl phosphite, the latter serving as both reductant and ligand.5 A 68% yield of a single crystalline (mp 85-86 "C) adduct is formed whose spectral properties identify it as the desired [6 + 31 adduct 3, R = R' = H.' The symmetry of the adduct is clearly visible by the simplicity of the ' H N M R spectrum [6 5.80 (td, J = 12.0, 3.1 Hz, 2 H), 5.42 (m, 2 H), 4.95 (appt., J = 1.5 Hz, 2 H), 3.37 (appt., J = 6.0 Hz, 2 H), 2.72 (ddd, J = 13.0, 6.0, 1.5 Hz, 2 H), 2.43 (d, J = 13.0 Hz, 2 H)]. The examples summarized in Table I attest to the generality of this reaction. To examine the role of the electrophilicity of the tropone, we placed electron-donating groups at the 2-position (1) Salem, L. J . Am. Chem. SOC.1968, 90, 543. Biggi, G.;Deloma, F.; Pietra, F. J . Am. Cfiem.SOC.1973, 95, 7101. For a review, see: Pietra, F. Acr. Chem. Res. 1979, 12, 132. (2) For anion additions to tropone and its use in natural products syntheses, see: Rigby, J.; Wilson, J. 2.J . Am. Chem. Soc. 1984, 106, 8217. (3) Ito, S.; Fujise, Y.; Okuda, T.; Inone, Y. Bull. Chem. SOC.Jpn. 1966, 39, 1351. Cookson, R. C.; Drake, B. V.; Hudec, J.; Morrison, A. Chem. Commun. 1966, 15. Ito, S.; Otani, H.; Narita, S.; Honma, H. Tetrahedron Lett. 1972, 2223. Mukai, T.; Akasaki, Y.; Hagiwara, T. J . Am. Chem. SOC. 1972, 94, 675. Garst, M. E.; Roberts, V. A,; Prussin, C. Tetrahedron 1978, 39, 581. Fujise, Y . ;Shiokawa, T.; Mazaki, Y . ;Fukazawa, Y.; Fujii, M.; Ito, S . Tetrahedron Letr. 1982, 23, 1601. Mukerjee, D.; Watts, C. R.; Houk, K. N. J. Org. Chem. 1983, 43, 817. Garst, M. E.; Roberts, V. A,; Houk, K. N.; Rondan, N. G. J . Am. Chem. SOC.1984,106, 3882. Paquette, L. A,; Hathaway, S. J.; Schirch, P. F. T. J . Org. Chem. 1985, 50, 4199. Rigby, J. H.; Moore, T. L ; Rege, S . J . Org. Chem. 1986, 5 1 , 298. Funk, R. L.; Bolton, G. L. J. Am. Chem. SOC.'986, 108, 4655. (4) Trost, B. M.; Chan, D. M. T . J . Am. Chem. Soc. 1983, 105, 2315, 2326. (5) Trost, B. M.; Nanninga, T. N.; Satoh, T. J . Am. Chem. SOC.1985, 107. 721. (6) Trost, B. M. Angew. Chem., Int. Ed. Engl. 1986, 25, 1 . For a review of a related reaction of methylenecyclopropanes, see: Binger, P.; Buch, H . M. Topics Curr. Chem. 1986, 135, 000. (7) All new compounds have been fully characterized and have had elemental compositions determined by combustion analysis and/or high-resolution mass spectroscopy.
0 1987 American Chemical Society
